286 Int. J. Environment and Pollution, Vol. 39, Nos. 3/4, 2009

Copyright © 2009 Inderscience Enterprises Ltd.

Atmospheric aerosols behaviour at an industrial area in Northern France

F. Ledoux* Université du Littoral Côte d’Opale, L.I.S.E., FRE 2816 ELICO, 32, avenue Foch, 62930 Wimereux, France E-mail: ledoux@univ-littoral.fr *Corresponding author

D. Courcot Université du Littoral Côte d’Opale, L.C.E., EA 2598, 145, avenue M. Schumann, 59140 Dunkerque, France E-mail: courcot@univ-littotal.fr

L. Courcot Université du Littoral Côte d’Opale, L.I.S.E., FRE 2816 ELICO, 32, avenue Foch, 62930 Wimereux, France E-mail: lucie@univ-littoral.fr

A. Aboukaïs Université du Littoral Côte d’Opale, L.C.E., EA 2598, 145, avenue M. Schumann, 59140 Dunkerque, France E-mail: aboukais@univ-littoral.fr

E. Puskaric Université du Littoral Côte d’Opale, L.I.S.E., FRE 2816 ELICO, 32, avenue Foch, 62930 Wimereux, France E-mail: puskaric@univ-littoral.fr

Abstract: Aerosols samplings by bulk filtration and by use of a high volume cascade impactor were performed taking into account wind directions simultaneously at an urban and a rural site in Northern France. Fifteen major and trace species were considered to study characteristics of the particles collected. Evolution of the Weighted Mean Diameter vs. the wind direction,

Atmospheric aerosols behaviour at an industrial area in Northern France 287

profiles of the Mass Size Function (MSF), values of the dry deposition fluxes, distribution between the coarse particles (>1 µm) and the fine particles (<1 µm), neutralisation ratio between NH4

+ and SO42−, NO3

− are discussed in this work.

Keywords: urban and rural aerosols; mass median diameter; mass size distribution; dry deposition fluxes; neutralisation ratio.

Reference to this paper should be made as follows: Ledoux, F., Courcot, D., Courcot, L., Aboukaïs, A. and Puskaric, E. (2009) ‘Atmospheric aerosols behaviour at an industrial area in Northern France’, Int. J. Environment and Pollution, Vol. 39, Nos. 3/4, pp.286–305.

Biographical notes: Frederic Ledoux gained his PhD in Chemistry at the ULCO in 2003. His researches deal with the characterisation of atmospheric particles in relation with the establishment of emission sources markers.

Dominique Courcot is Professor at the University of Littoral Côte d’Opale (ULCO, France). His researches lie in the fields of environmental sciences, and more specifically on the physicochemical characterisation of atmospheric particulate matter. An other aspect of his studies is devoted to the properties of some metal oxides as potential catalysts for the elimination of vehicles exhausts.

Lucie Courcot is an Engineer in analytical chemistry expert on environmental topics. She is particularly involved in studies dealing with the application of Scanning Electron Microscopy for the characterisation of atmospheric aerosols and biological organisms present in sea water.

Antoine Aboukaïs is the Director of the Laboratoire de Catalyse et Environnement linked to the University of Littoral Côte d’Opale. He is working on elimination of VOCs, combustion of soots, characterisation of atmospheric aerosols and toxicology.

Emile Puskaric is Professor at the University of Littoral Côte d’Opale, and his first research during several years’ interests inorganic products synthesis before studying fluxes and chemical properties of atmospheric trace metals in sea water. Since about 15 years, he studies particularly the characterisation of the atmospheric aerosols components produced by the industrial activities and their influence on the environment, in particular in the field of toxicology.

1 Introduction

Nord – Pas de Calais in Northern France, is one of the most industrialised countries of France, and in 2001, it was estimated that it was responsible of 7000 tons of atmospheric particulate matter emission. In this region, 60% of this kind of emissions is originated from the high-industrialised Dunkerque area. If a great number of studies have been made especially on the Belgian coast of the North Sea among which

288 F. Ledoux et al.

Bruynseels et al. (1988), Injuk et al. (1992) and Rojas et al. (1993) can be cited, few of them concern this kind of work on this French coast (Flament et al., 1987). The aim of this work is the characterisation of particulate aerosols in a heavy industrialised area such as Northern France, counting a lot of heavy metals emission sources (metallurgic industries, smelters, waste incinerators …). The present study attempts to determine the chemical composition of particles in selected size fractions according to source contribution and the behaviour of some elements in the atmosphere.

2 Methodology

2.1 Study area

Aerosol samples were collected from 15th January 2001 to 23rd February 2001. The sampling sites are located at Dunkerque an important urban city with roughly 220,000 inhabitants with its suburbs, and where important industrial activities are concentrated, and at a rural area, Les Moëres, a village with 860 inhabitants, situated approximately 15 km ESE of Dunkerque, as shown in Figure 1.

Figure 1 Study area with location of the sampling sites

The sampling site at Dunkerque is under the important influence of numerous aerosols sources, as urban origins, as well as particularly westside industrial sources (metallurgical plants, oil refineries, organic chemistry), and also marine sources under the northside marine influence of the North Sea. Three wind sectors are taken into account from the different wind directions with respectively the marine sector (340–60°), the continental and urban sector (60–240°) and the industrial sector (240–340°). For the rural sampling site, these three wind sectors are also used, with the difference that the 60–240° sector is only under the continental influence, and the 240–340° sector is under the complete influence of Dunkerque area.

Atmospheric aerosols behaviour at an industrial area in Northern France 289

2.2 Sampling methods

2.2.1 Bulk filtration

The total atmospheric particulate matter, 74 samples at the urban site and 74 samples at the rural site, is collected by bulk filtration using cellulose filters (Whatman 41, porosity 0.04 µm) as filtration media. The filter is mounted in a Polyvinyl Chloride (PVC) holder placed at the top of a 10 m mast ended by a 1 m PVC head to avoid any metallic contamination from the mast. An unoiled pump operating at a 5 m3.h–1 flow rate is connected to the sampling head by a 20 mm i.d. polyethylene tube and allows to filter a volume of 60 m3 of air for each sample. The sampling head is changed every 12 hours (at 8 pm and 8 am) and after each sampling time, the filter is picked up from the holder under a laminar flow hood, put in a clean Petri box and stored in a freezer (–20°C) until the analysis to avoid, as much as possible, any volatilisation of nitrate and chloride. A blank filter in its holder is put also at the top of the mast but without running. A meteorological station was set up at each sampling site to obtain local wind speed and direction, atmospheric pressure, temperature, hygrometry, rain fall.

2.2.2 Cascade impaction sampling

Size-fractionated atmospheric particulate matter are sampled on Whatman® 41 cellulose filters by use of a 5 stage plus back-up filter high volume cascade impactor (Sierra Instruments, Model 235). For each impaction stage, experimental values for the aerodynamic 50% cut-off particle diameter (Dp,50) as recommended by Willeke (1975) and used by Baeyens et al. (1990) and Flament et al. (1996) in similar studies with same collection substrat and nominal flow rate of 68 m3.h–1, are indicated in Table 1.

Table 1 Calculated 50% cut-off particle diameters (Dp,50) under a nominal flow rate of 68 m3.h–1 for the different stages of the cascade impactor Sierra Model 235

Impactor stage Dp,50 (µm) 1 5.08 2 2.10 3 1.04 4 0.64 5 0.33 6 0.04

The particles were divided in coarse particles (>1 µm) sampled on the stages 1, 2 and 3, and fine particles (<1 µm) collected on the stages 4, 5 and 6 (back-up).

2.3 Filters preparation

Unwashed Whatman®41 cellulose filters used for bulk filtration, present varying trace metals contents and especially for Fe, up to 5 µg per filter. So this kind of filters needs to be washed before sampling and they were treated as described by Obata et al. (1997) for Fe analysis. Filters are heated at 80°C in a 0.5 M aqueous solution of HCl, HNO3 and HF (Suprapur® Grade, Merck Darmstad) during 1 h, then dived into HCl 1 M

290 F. Ledoux et al.

and ascorbic acid (0.1 M) during 24 h, rinsed with HCl 1 M and finally with ultrapure water (Milli Q®, Millipore, resistivity 18.2 MΩ.cm). Filters are then dried under a laminar flow hood (Class 100, US Federal Standard 209a), and then hermetically kept in cleaned Petri boxes and stored in a freezer (–20°C). Such a treatment lowers the Fe contamination under 0.1 µg per filter, and contents of all other elements are lower than detection limit.

2.4 Chemical analysis

Fifteen elements and ions have been taken into account for this study, such as major components (Cl–, NO3

–, SO42–, NH4

+, Na, Mg and Ca), and trace elements (Al, Cu, Fe, K, Mn, Pb, Sr and Zn).

2.4.1 ICP-AES and GFAAS method

Inductively Coupled Plasma – Atomic Emission Spectrometry (ICP-AES, Varian, Liberty II) was used to quantify Al, Ca, Fe, K, Mg, Mn, Na, Sr, and Zn, whereas Cu and Pb concentrations were determined with the Graphite Furnace Atomic Absorption Spectrometry (GFAAS) technique (Perkin Elmer, AAnalyst 600, Zeeman background correction) after acid digestion of the samples. A part of each exposed or unexposed cellulose filter was placed into a PTFE vessel (SAVILLEX® Corp.) and digested with 6 ml HNO3/HF/HClO4, (4 : 1 : 1 v/v, Suprapur® , Merck®) at 130°C for 4 h. The acids were then evaporated at 170°C for 4 h using hot plates, and ultrapure water was added to the residue and heated at 110°C during 1 h. Finally, the cooled solution was diluted with MilliQ® water up to 10 ml in presence with 0.2% of nitric acid, for analysis.

2.4.2 Ion Chromatography and colorimetric method

Cl–, NO3–, SO4

2– and NH4+ ions were quantified with ion chromatography or colorimetric

method after water extraction from the atmospheric particles. A part of each cellulose filter was placed into a glass vessel with 4 mL milliQ water. Cl–, NO3

–, SO42– and NH4

+ were extracted from the filter by leaching with a 30 min ultrasonic treatment. This step was repeated three times with fresh water and for each filter the solutions were finally gathered, filtered on 0.2 µm cellulose acetate membrane and the volume adjusted up to 20 mL. The Dionex DX 500 ion chromatography, equipped with electrochemical suppressor and conductivity detector was used with a NaHCO3 0.5 mM and Na2CO3 4.5 mM mobile phase for quantification of Cl–, NO3

– and SO42–. NH4

+ concentrations were determined using the indophenol-blue colorimetric method. This method is based on the formation of an indophenol blue compound by reaction between NH4

+ ions with phenol and sodium hypochlorite using nitroprussiate as catalyst. After a 12 h reaction period, absorbance is determined using a spectrometer at 630 nm and is proportional to the NH4

+ content.

2.4.3 Accuracy and validity of the quantification methods

For ICP-AES, GFAAS, ion chromatography and colorimetric method, measurements were repeated three times for all samples and the concentration attributed to the sample corresponds to the mean value. For a considered sample, the maximum standard

Atmospheric aerosols behaviour at an industrial area in Northern France 291

deviation observed was less than 3%. The quantification methods have been also validated using the Standard Reference Material “urban particulate matter 1648” of NIST. Differences obtained between measured concentrations and certified values were less than 4%. Thus, we can conclude that the methods used are accurate and validated for Al, Ca Cu, Fe, K, Mg Mn, Na, Pb, Sr, Zn and Cl–, NO3

– SO42– and NH4

+ quantification in atmospheric aerosols.

3 Results and discussion

3.1 Elemental and ionic concentrations in the atmospheric particulate matter

Geometric and arithmetic mean elemental and ionic concentrations calculated from the results obtained for the bulk filtration samples, are given in Table 2 for the two sites and for the whole sampling period. The mean concentrations observed at the urban site are all higher than those than at the rural site which evidenced very well the influence of the urban and the industrial activities on the atmospheric particles composition. Nevertheless for some elements the difference between the two sites mean concentration values is more important than for others, in particular for Mn and Fe. For these elements, concentrations are respectively 3.4, 3.3 times higher at the urban site than at the rural site whereas for Cu and Ca, the ratio between urban and rural site concentrations is 2.8 and 2.4 times and for the other elements and species this ratio is less than 2. These observations lead to underline that the industrial park has a role in the chemical composition of the atmospheric particles at the urban site. In fact, Mn and Fe elements can be easily linked to the ferromanganese producer and to the steel metallurgy plant implanted in Dunkerque area. Differences observed for other elements could be explained on the one hand by some other industrial and human activities and on the other hand by the proximity of the sea in particular for Cl–, Na and Mg.

Table 2 Average concentrations (ng.m–3) of the elements and ions in atmospheric particles sampled at the urban and rural sites. µG and µA are respectively geometric and arithmetic means, σ is the standard deviation

Urban site Rural site

µG µA σ µG µA σ Al 159 274 373 88 133 110 Ca 591 1121 1509 248 370 294 Cu 8.9 17 23 3.2 6.0 6.6 Fe 475 962 1174 143 296 329 K 252 388 459 161 236 217 Mg 170 252 204 113 163 146 Mn 27 65 94 8.1 24 42 Na 916 1561 1730 641 1053 1114 Pb 26 56 80 15 31 33 Sr 2.4 3.7 3.9 1.3 1.7 1.1

292 F. Ledoux et al.

Table 2 Average concentrations (ng.m–3) of the elements and ions in atmospheric particles sampled at the urban and rural sites. µG and µA are respectively geometric and arithmetic means, σ is the standard deviation (continued)

Urban site Rural site

µG µA σ µG µA σ Zn 61 164 333 34 90 134 Cl– 2343 3427 3466 1577 2151 1629 NO3

– 4027 6531 6560 3467 5503 4580 SO4

2– 3051 4453 4258 2291 3629 3391 NH4

+ 2532 3871 3763 2177 3120 2382

3.2 Weighted Mean Diameter (WMD)

The mean size of particles associated with an element or ions (x) is determined by the ‘Weighted Mean Diameter’ (WMD), usually expressed in µm. More details regarding the calculation of the WMD can be found in Flament et al. (1987, 1996) and Véron et al. (1999). By definition:

6,50=1

6

=1

WMD( )=i

i pi

ii

C xDx

C∑∑

where Ci is the elemental concentration on the ith stage of impaction (ng.m–3) and ,50ipD

the corresponding 50% cut-off particle diameter. The evolution of the behaviour of the different elements and ions was followed by

their calculated WMD according to the wind directions, shown in Figure 2. Elements and ions with similar profiles are gathered on the same graph, but Cu, K, Pb

and Zn with their particular profiles cannot be superimposed with the other elements. Na and Cl are shown separately considering their affinity when originated from the marine sector. At the both sampling sites, the size of the particles related to these two elements are very well correlated for the marine sector (340–60°), on the other hand these particles present different profiles especially for the 60–240° sector which is strictly continental for the rural site, and continental with urban area influence for the urban site. The WMD of the particles associated with chloride mainly lower than those associated with Na, evidence an anthropogenic contribution to the presence of Cl– in the atmospheric particles.

At the urban site, three groups of elements with identical profiles are identified:

• 23 4 4NO , SO and NH− − +

• Al, Ca, Mg and Sr

• Fe and Mn.

At the rural site, these elements are gathered in exactly the same way as at the urban area, but with notably differences for the 240–340° sector, corresponding to the important influence of the industrialised area of Dunkerque, especially for NO3

–, Al and Ca.

Atmospheric aerosols behaviour at an industrial area in Northern France 293

Figure 2 Evolution of Weighted Mean Diameters (WMD) for elements and ions at each sampling site vs. the wind direction: (a) urban site and (b) rural site

(a) (b)

294 F. Ledoux et al.

3.3 Mass Size Function (MSF)

The distribution of an element or ion (x) with respect to particle size is expressed in terms of a discrete mass-size distribution, on a log-scale, called the Mass Size Function (MSF). The MSF is obtained by plotting the mass percentage on each stage of impaction vs. the aerodynamic 50% cut-off particle diameter Dp,50 (Friedlander, 1977). By definition:

,501,50 ,50

/MSF( ) (log )

(log ) (log )ii Tpi i

p p

C Cx f D

D D−= =−

where Ci is the concentration of the element (x) (ng.m–3) at the stage i, and CT the total concentration of this element during the sampling. CT = C1 + C2 + + C6.

Tables 3 and 4 present the distribution of fine and coarse particles for the sampled elements according to the different wind sectors at the both sampling sites. The 340–60° wind sector represents correctly the marine sector for the sampling sites with Ca, Mg, Na, Sr and Cl which are in majority coarse particles (>1 µm), with the higher percentage in this mode whatever the sector may be for Mg, Na and Sr, and also for Al at the rural site. On the other hand, for the urban site, the other trace elements have their main percentage in the fine mode (<1 µm) especially in the continental sector (60–240°) under the important influence of the city and in the industrial sector (240–340°), source of main anthropogenic elements. This mode is also found at the rural site for the trace elements except for Al and Fe.

The MSF allows us to determine the behaviour of the different aerosols according to their size distribution, and it is well known, as described by Safai et al. (1993), that this size distribution is generally similar and often bimodal with differences in the percentage contributions of particles present in the two modes (Horvath et al., 1996). The MSF of the sampled elements at the both sampling sites, for a precise wind direction at each sector, are presented respectively for the major elements in Figure 3 and for the trace metals in Figure 4. The major elements are characterised, as shown in Figure 3, by an unimodal distribution for both sampling sites for the marine sector, and these ions dominate in the coarse mode (2.1–5.08 µm) except +

4NH with a peak in the fine mode (0.64–1.04 µm), and 2-

4SO having a bimodal distribution for the urban site with a peak between 0.33 and 0.64 µm.

All these major aerosols have a bimodal distribution in the continental sector for the urban site, and only Ca, Cl– and 3NO− have the same kind of distribution for the rural site. It can be noted that the elements characterised by a bimodal distribution have common peaks between 0.64–1.04 µm and 2.1–5.08 µm, but Cl–, 3NO ,− 2

4SO − and +4NH from the

rural site, present a peak between 0.33–0.64 µm instead of 0.64–1.04 µm. This difference between the two sampling sites results probably from the importance of the influence of the urban area on the urban site. For the industrial sector, the same peaks are always present for the bimodal distribution, with the difference that for the rural site, Cl–, 3NO− and 2

4SO − have a peak in the 2.1–5.08 µm range and 4NH+ with a unimodal distribution, dominates with a peak at 0.64 and 2.1 µm.

Atmospheric aerosols behaviour at an industrial area in Northern France 295

Table 3 Elements and ions composition of size-resolved particles at the urban site according to wind sectors (values in percent)

296 F. Ledoux et al.

Table 4 Elements and ions composition of size-resolved particles at the rural site according to wind sectors (values in percent)

Atmospheric aerosols behaviour at an industrial area in Northern France 297

Figure 3 Mass Size Function of major elements at the sampling sites according to wind sectors (Grey colour: Urban Site; Blue colour: Rural Site) (see online version for colours)

Taking into account the trace elements, we see in Figure 4, that their distribution can be unimodal or bimodal according to wind sectors. For the marine sector, Al, Mn, Pb, Sr and Zn at the urban site, and K and Sr at the rural site, present a unimodal distribution. Al, K and Sr dominate in the coarse mode with diameters of the particles between 2.1 µm

298 F. Ledoux et al.

and 5.8 µm. The diameters of the particles including other elements, are between 0.33 and 0.64 µm and 0.64 and 1.04 µm respectively and are included in a wider range for Zn (0.33–1.04 µm).

Figure 4 Mass Size Function of minor elements at the sampling sites according to wind sectors (Grey colour: Urban Site; Blue colour: Rural Site) (see online version for colours)

Atmospheric aerosols behaviour at an industrial area in Northern France 299

For the continental and industrial sectors, all these trace elements present a bimodal distribution at the urban site. In that case, the diameters of the particles in the two modes are between 0.64 and 1.04 µm and 2.1 and 5.08 µm. The fine mode is the main result of the influence of the industrial area, and this is particularly shown by the slipping peak from 0.64–1.04 µm to 0.33–0.64 µm for Pb and above all K and Zn. For the same sectors at the rural site, the diameters of the particles are equivalent in the two modes, very important between 2.10 µm and 5.08 µm for all elements, except Zn at the continental sector. The influence of the industrial sector appears with most particles submicrometer sized (0.64–1.04 µm).

3.4 Assessment of dry deposition fluxes

Important dry removal mechanisms for scavenging aerosols from the atmosphere are impaction, diffusion and sedimentation, which are dependent upon wind speed (Hussein et al., 2006). Under dry conditions, sedimentation determines the upper limit of the aerosol size distribution, and the ground dry deposit is usually described as a dry removal velocity, which corresponds to the rate between the particle settling by unit of surface and the concentration in the atmosphere, and as we know, the gravitational settling of the particles highly depends of their size, their composition and weather conditions.

3.4.1 Determination of the deposition velocity

The calculation of the dry deposition fluxes of particulate matter can be done as presented by Duce et al. (1991), who take into account settling velocity of the particles according to the Enrichment Factor (EF) with V = 0.1 cm.s–1 for elements with EF < 10 and V = 1 cm.s–1 for elements with EF > 10. On the other hand, Golomb et al. (1997) attribute a settling velocity (V = 0.025 cm.s–1) for fine particles (<2.5 µm) and a settling velocity of 0.25 cm.s–1 for the coarse particles (>2.5 µm).

One of the most frequently used model is the Slinn and Slinn (1980) model where the different kinds of transport (turbulent diffusion, Brownian motion), impaction opportunity and size increasing particles near the interface, are taken into account. This model allows the gravitational settling velocity of the particles according to their size and takes also into account the variations of the settling velocity according to their aerodynamic diameter. We have chosen this model because it is usually used for our kind of study, for sampling sites near sea-side like Dulac et al. (1989), Baeyens et al. (1990), Bertho et al. (1998) and Injuk et al. (1998).

The deposition velocity can be determined from the application of the Slinn and Slinn (1980) model, as presented in Figure 5 and using the WMD values of the components as determined previously.

The deposition velocities presented in Table 5, are relatively homogenous for the different elements and are equivalent for each sampling site, between 0.19 and 0.63 cm.s–1. Pb presents the lowest velocity (0.19 cm.s–1) and is the same as determined at Cape Gris-Nez (Bertho et al., 1998), a site about 50 km NW apart our studied area.

On the other hand, we can observe that our deposition velocities agree the hypothesis of Duce et al. (1991) with velocities between 0.1 cm.s–1 and 1 cm.s–1 for non enriched and enriched elements respectively, as seen previously.

300 F. Ledoux et al.

Figure 5 Deposition velocity (Vd) in dependence with the particle diameter

Source: From Slinn and Slinn (1980)

Table 5 Weighted Mean Diameter (WMD) and corresponding deposition velocities (Vd) determined from the Slinn and Slinn model, for each sampling site – σWMD and σVd are respectively the standard deviation on WMD and Vd values

Urban site Rural site

WMD (µm) σWMD

Vd (cm.s–1) σVd

WMD (µm) σWMD

Vd (cm.s–1) σVd

Al 1.40 0.36 0.50 0.07 1.60 0.42 0.55 0.06 Ca 1.52 0.39 0.52 0.05 1.66 0.35 0.56 0.05 Cu 1.12 0.17 0.42 0.06 1.23 0.52 0.46 0.10 Fe 1.42 0.44 0.50 0.08 1.43 0.48 0.50 0.08 K 1.09 0.56 0.41 0.15 1.30 0.60 0.47 0.17 Mg 1.80 0.52 0.58 0.06 1.77 0.50 0.58 0.06 Mn 1.21 0.39 0.44 0.11 1.19 0.56 0.44 0.19 Na 2.09 0.42 0.63 0.03 2.06 0.49 0.62 0.02 Pb 0.69 0.26 0.19 0.09 0.68 0.41 0.19 0.17 Sr 1.67 0.47 0.56 0.06 1.79 0.42 0.57 0.06 Zn 0.96 0.31 0.35 0.09 1.05 0.51 0.39 0.16 Cl– 1.91 0.67 0.60 0.08 1.71 0.87 0.57 0.18 NO3

– 0.90 0.57 0.31 0.16 0.89 0.62 0.31 0.21 SO4

2– 0.85 0.49 0.27 0.19 0.87 0.42 0.29 0.19 NH4

+ 0.72 0.32 0.20 0.15 0.72 0.32 0.20 0.17

3.4.2 Total dry deposition fluxes

The dry deposition fluxes (Fdry) are calculated using the equation:

Fdry = 0.32 × C × Vd

where Fdry is the dry deposition (kg.km–2.y–1), C the geometric mean concentration of the components (ng.m–3), Vd the elemental settling velocity (cm.s–1) and 0.32 is the coefficient obtained according to the different unities. Geometric mean is advocated by

Atmospheric aerosols behaviour at an industrial area in Northern France 301

several authors (Chester and Bradshaw, 1991; Van Jaarsveld, 1993; Ottley and Harrison, 1993; Rojas et al., 1993) with respect to arithmetic mean, because this last calculation provides a stronger statistical importance to episodically measured high values, not representative.

Our results are compared with those of other authors in Table 6. The comparison with the results obtained by Flament (1985) is very interesting because they were issued from sampling sites near our urban area, Dunkerque, and we can see in Table 6, that our results are for the most part, lower. This evolution can be explained by the more and more strict environmental measures taken against industrial emissions.

Table 6 Elements and ions dry deposition fluxes (kg.km–2.y–1) determined at the urban and rural sites and in some European locations

Flament (1985)

Injuk et al. (1998)

Chester et al. (1999) This study (2001)

This study (2001)

Sampling site North Sea

French coast North Sea

Belgian coast Mediterr. coastDunkerque (Urban site)

Les Moëres (Rural site)

Al 82.7 139 25.5 15.5 Ca 25 98.4 44.5 Cu 4.1 0.5 0.2 1.2 0.5 Fe 133.7 24 103 76.1 22.9 K 75.5 16 33.1 24.3 Mg 31.7 21.0 Mn 13.7 0.8 3.9 3.9 1.1 Na 379 185 127 Pb 11.2 0.4 2.11 1.6 0.9 Sr 0.4 0.3 Zn 36.3 2 1.6 6.8 4.2 Cl– 450 288 NO3

– 400 344 SO4

2– 264 213 NH4

+ 162 139

The more significant result concerns Pb for which the flux, since 1985, is divided by a factor 7 and about 12, at the urban site and the rural site respectively thanks to the environmental measures and the decrease of the automotive lead. Except this consideration, our results at the urban site are all higher compared to the other works. By comparison with the data of Injuk et al. (1998) and concerning a study area near our urban sampling site, we see clear differences for all elements and particularly for Fe, Mn, Pb and above all Zn. This shows very well the influence of the important industrial activities in this studied area. As we have seen previously, the high values for Zn are not only due to the industrial influence, but are, above all, originated from the urban activity. The noted fluxes for this trace element are very high as well at the urban area as at the rural site, respectively 6.8 and 4.2 kg.km–2.y–1, and much higher than those observed on the North Sea Belgian coast by Injuk et al. (1998) or on the western Mediterranean coast, by Chester et al. (1999).

302 F. Ledoux et al.

3.5 Neutralisation of particles

From the analytical results, it appears that most molar ratios of 24 4NH / SO ,+ − 4 3NH / NO+ −

and 4NH / Cl+ − are widely higher than theoretical values which are respectively 2, 1 and 1. This shows that a noticeable amount of 4NH+ is combined with organic compounds as well with coarse particles as with fine particles. In coarse particles, the molar ratio 4NH / Cl+ − is often below 1, meaning that ammonium chloride is also present with coarse particles as explained by Zuhang et al. (1999). This observation easily allows the explanation of Neutralisation Ratios (NR) obtained during the sampling campaign, using the following formula:

+ 34

2 3 34 3

NH (nmol.m )NR = (nss = non-sea-salt).2.nssSO (nmol.m ) + NO (nmol.m )

−

− − − −

The amounts of 4NH+ are much higher in the submicronic field, and this ratio increases when the particles are in the fine mode. Figure 6 shows that, for fine particles, this ratio is higher than 1 at the both sampling sites, and in this case, it can be supposed the presence of NH4Cl or other forms of 4NH+ compounds, because lowest temperatures could stabilise this compound and limit its volatilisation.

Figure 6 Number of mole (n) of NH4+ vs. the sum (2n(SO4

2–) + n(NO3–)) for the determination

of the Neutralisation Ratios (NR) at each sampling site

On the other hand, a ratio close to 1, indicates that 4NH+ is mainly present as (NH4)2SO4 and NH4NO3. For each sampling site this ratio for coarse particles is lower than 1 and indicates that 2

4SO − and 3NO− are associated with other elements like Na and Ca (Harrison et al., 1994; Wakamatsu et al., 1996).

Atmospheric aerosols behaviour at an industrial area in Northern France 303

4 Conclusion

Weighted Mean Diameter reveals that the elements have different profiles according to the wind sectors, and for each sampling site, components presenting industrial profiles can be classified in three groups:

• 23 4 4NO , SO , NH− − +

• Al, Ca, Mg, Sr

• Fe and Mn.

The influence of the industrialised area on the rural site appears very well because of differences in the profiles of some trace elements. The MSF study shows for most of the elements a bimodal mode, corresponding to several sources. One peak occurred between 0.33–0.64 µm or 0.64–1.04 µm, and the second between 2.10–5.08 µm. Aerosols exist in the fine and the coarse mode, but differ in percentage distribution according to the origin of the winds. Anthropogenic elements and fine particles, are encountered for the industrial wind sector, and coarse particles, notably major elements, characterise the marine wind sector. 2

3 4 4NO , SO , and NH ,− − + at the urban sampling site, are mainly originated from gas-particle conversion because they present a unimodal distribution especially in the fine mode. The study of the dry deposition fluxes shows very well the concentration decreasing of most components compared with results of other authors and the decrease of Pb in the atmosphere reflects perfectly the present tendency of the environmental measures. According to the importance of the neutralisation ratio study between 4NH+ and 2

4SO − + 3NO ,− the behaviour of 4NH+ in the aerosols can be explained.

Acknowledgements

The authors wish to acknowledge support for this work from Nord-Pas de Calais Council (No. 02050163) and FEDER Grant (No. 337/777).

References Baeyens, W., Dehairs, F. and Dedeurwaerder, H. (1990) ‘Wet and dry deposition fluxes above the

North Sea’, Atmospheric Environment, Vol. 24A, No. 7, pp.1693–1703. Bertho, M.L. (1998) Le plomb associé aux aérosols atmosphériques dans le Détroit du Pas de

Calais: sources, flux et devenir à l’interface air-mer, PhD Thesis, University of Littoral Côte d’Opale, France, p.275.

Bruynseels, F., Storms, H. and Van Grieken, R. (1988) ‘Characterisation of North Sea aerosols by individual particle analysis’, Atmospheric Environment, Vol. 22, pp.2593–2602.

Chester, R. and Bradshaw, G.F. (1991) ‘Source control on the distribution of particulate trace metals in the North Sea atmosphere’, Mar. Poll. Bull., Vol. 22, pp.30–36.

Chester, R., Nimmo, M. and Preston, M.R. (1999) ‘The trace metal chemistry of atmospheric dry deposition samples collected at Cap Ferrat: a coastal site in the western Mediterranean’, Marine Chemistry, Vol. 68, pp.15–30.

304 F. Ledoux et al.

Duce, R.A., Liss, P.S., Merril, J.T., Atlas, P., Buat-Menard, P., Hicks, B.B., Miller, J.M., Prospero, J.M., Arimoto, R., Church, T.M., Ellis, W., Galloway, j.N., Hansen, I., Jickells, T.D., Knap, A.H., Reinhard, K.H., Schneider, B., Saoudine, A., Tokas, J.J., Tsungogai, S., Wollast, R. and Zhou, M. (1991) ‘The atmospheric inputs of trace species to the World Ocean’, Global Biogeochem. Cycles, Vol. 5, pp.193–259.

Dulac, F., Buat-Menard, P., Ezat, U., Melki, S. and Bergametti, G. (1989) ‘Atmospheric input of trace metals of the western Mediterranean uncertainties in modelling dry deposition from cascade impactor data’, Tellus, Vol. 41B, pp.362–378.

Flament, P. (1985) Les métaux-traces associés aux aérosols atmosphériques: apports au milieu marin du littoral Nord-Pas de Calais, PhD Thesis, University of Lille, France, p.175.

Flament, P., Leprêtre, A. and Noël, S. (1987) ‘Aérosols côtiers dans le nord de la Manche’, Oceanologica Acta, Vol. 10, pp.49–61.

Flament, P., Bertho, M.L., Deboudt, K. and Puskaric, E. (1996) ‘Changes in the lead content of atmospheric aerosols above the Eastern Channel between 1982/83 and 1994’, The Science of the Total Environment, Vol. 192, pp.193–206.

Friedlander, S.K. (1977) Smoke, Dust and Haze: Fundamentals of Aerosol Behaviour, Wiley, New York, p.317.

Golomb, D., Ryan, D., Eby, N., Underhill, J. and Zemba, S. (1997) ‘Atmospheric deposition of toxics onto Massachussets bay. I – Metals’, Atmospheric Environment, Vol. 31, pp.1349–1359.

Harrison, R.M., Msibi, M.I., Kitto, A.M.N. and Yamulki, S. (1994) ‘Atmospheric chemical transformations of nitrogen compounds measured in the North Sea experiment, September 1991’, Atmospheric Environment, Vol. 28, pp.1593–1599.

Horvath, H., Kasahara, M. and Pesava, P. (1996) ‘The size distribution and composition of the atmospheric aerosol at a rural and nearby urban location’, Journal of Aerosol Science, Vol. 27, pp.417–435.

Hussein, T., Karppinen, A., Kukkonen, J., Härkönen, J., Aalto, P.P., Hämeri, K., Kerminen, V.M. and Kulmala, M. (2006) ‘Meteorological dependence of size-fractionated number concentrations of urban aerosol particles’, Atmospheric Environment, Vol. 40, pp.1427–1440.

Injuk, J., Otten, P., Laane, R., Maenhaut, W. and Van Grieken, R. (1992) ‘Atmospheric concentrations and size distribution of aircraft-sampled Cd, Cu, Pb and Zn over the southern bight of the North Sea’, Atmospheric Environment, Vol. 26A, pp.2499–2508.

Injuk, J., Van Grieken, R. and De Leeuw, G. (1998) ‘Deposition of atmospheric trace elements into the North Sea: coastal, ship, platform measurements and model predictions’, Atmospheric Environment, Vol. 32, pp.3011–3025.

Obata, H., Karatani, H. and Nakayama, E. (1997) ‘Fundamental studies for chemical speciation of iron in seawater with an improved analytical method’, Marine Chemistry, Vol. 56, pp.97–106.

Ottley, C.J. and Harrison, R.M. (1993) ‘Atmospheric dry deposition flux of metallic species to the North Sea’, Atmospheric Environment, Vol. 27A, pp.685–695.

Rojas, C.M., Injuk, J., Van Grieken, R. and Laane, R.W. (1993) ‘Dry and wet deposition fluxes of Cd, Cu, Pb and Zn into the southern bight of the North Sea’, Atmospheric Environment, Vol. 27A, pp.251–259.

Safai, P.D., Khemani, L.T., Momin, G.A., Rao, P.S.P. and Pillai, A.G. (1993) ‘Mass size distribution and chemical composition of aerosols in the Silent Valley, India’, Indian Journal of Radio and Space Physics, Vol. 22, pp.56–61.

Slinn, S.A. and Slinn, W.G.N. (1980) ‘Prediction for particle deposition on natural waters’, Atmospheric Environment, Vol. 14, pp.1013–1016.

Van Jaarsveld, J.A. (1993) ‘Atmospheric inputs to the North Sea and coastal waters’, J. IWEM, Vol. 7, pp.24–31.

Atmospheric aerosols behaviour at an industrial area in Northern France 305

Véron, A.J., Flament, P., Bertho, M.L., Alleman, L., Flegal, R. and Hamelin, B. (1999) ‘Isotopic evidence of pollutant lead sources in north-western France’, Atmospheric Environment, Vol. 33, pp.3377–3388.

Wakamatsu, S., Utsunomiya, A., Suk Han, J., Mori, A. and Uno, I. (1996) ‘Seasonal variation in atmospheric aerosols concentration covering Northern Kyushu, Japan and Seoul, Korea’, Atmospheric Environment, Vol. 30, pp.2343–2354.

Willeke, K. (1975) ‘Performance of the slotted impactor’, Am. Ind. Hyg. Assoc. J., Vol. 36, pp.683–691.

Zuhang, H., Chan, C.K., Fang, M. and Wexler, A.S. (1999) ‘Size distribution of particulate sulfate, nitrate and ammonium at a coastal site in Hong Kong’, Atmospheric Environment, Vol. 33, pp.843–853.